Image sensors are semiconductor devices having the function of photographing images of human beings or objects. The market of these image sensors has been rapidly expanded as they have been loaded into portable phones as well as common digital cameras or camcorders.
Such an image sensor is configured in the form of a camera module and mounted in the aforementioned apparatuses. A camera module comprises a lens, a holder, an infrared (IR) filter, an image sensor, and a printed circuit board. An image is formed by the lens of the camera module, the image formed by the lens is concentrated on the image sensor through the IR filter, and an optical signal of the image is converted into an electrical signal by the image sensor so as to photograph the image.
Among these components, the image sensor for converting an optical signal into an electrical signal is directly mounted as a bare chip on the camera module, or mounted on the camera module after an image sensor chip is packaged.
Among several methods of directly mounting a bare chip of an image sensor on a camera module, a COB (Chip-On board) method which currently occupies 90% or more has problems such as low productivity caused by a unit level packaging scheme, a high defective rate caused by introduction of dust particles during fabrication processes, high investment and maintenance costs of equipment including a clean room having a high degree of cleanness, and limitations on miniaturization. That is, all color filters and micro-lenses are very vulnerable to introduction of dust particles or penetration of moisture because they are fabricated through a photolithographic process after being coated with photoresist. Therefore, according to the COB method, the mounting of the image sensor chip, a wiring operation, the installation of the IR filter, lens and holder, and the like should be carried out in a clean room in which a high degree of cleanness is maintained.
On the contrary, if an image sensor that has been packaged in advance is used, it is possible to solve the aforementioned problems caused when the bare chip is used.
FIG. 1 shows a schematic sectional view of a ceramic leadless chip carrier (CLCC) most frequently used as an image sensor package. In a conventional image sensor package 20 shown in the figure, an image sensor chip 22 is mounted on a ceramic substrate 24 by using epoxy or the like such that the surface thereof faces upward, and the image sensor chip is then covered with a glass cover or glass substrate 21. In order to connect the image sensor chip 22 to the ceramic substrate 24, wires 26 connected to the image sensor chip 22 are connected to connection terminals 27 formed on a floor of the ceramic substrate 24, and the image sensor package 20 is connected to a circuit board by the connection terminals 27.
Another package method is to apply a chip scale package scheme (CSP) to an image sensor chip. This method allows an image sensor chip to be packaged at a wafer level contrary to the chip-on board (COB) method in which an image sensor chip as a bare chip is mounted on a camera module, thereby preventing dust or moisture from penetrating into an image sensing area.
An image sensor package 30 shown in FIG. 2 has been proposed by Schellcase Inc. Specifically, an image sensor chip 32 of which a bottom surface is polished to a thickness of about 100 micrometers is first prepared, an adhesive such as epoxy is coated to form an adhesive layer 34 on a top surface of the image sensor chip with a circuit formed thereon, a glass substrate 31 is then attached to the adhesive layer, an adhesive such as epoxy is then coated to form an adhesive layer 33 on the polished bottom surface, and a glass wafer 35 is then attached to the formed adhesive layer. Then, a dicing blade having a slightly gentle tip angle is used to remove a region between the image sensor chip 32 and the adhesive layer 34, thereby exposing input/output pads of the circuit formed on the top surface of the image sensor chip 32. Further, lateral sides of the image sensor chip 32, adhesive layer 33, and glass wafer 35 are formed to be inclined at a certain angle by using equipment such as a semiconductor wafer cutter (dicing saw). Next, metal wires 36 are formed to extend from the input/output pads of the exposed image sensor chip 32 via the inclined lateral side surfaces to a bottom surface of the glass wafer 35. At this time, the metal wires 36 are formed by forming a metal film from the input/output pads of the exposed image sensor chip 32 via the inclined lateral side surfaces to the bottom surface of the glass wafer 35 and by etching the metal film to form a desired pattern. Finally, connection terminals 37 such as solder balls are formed at ends of the metal wires 36 formed on the bottom surface of the glass wafer 35. The connection terminals 37 will be connected to external terminals or a printed circuit board (PCB). Such an image sensor package available from Shellcase Inc. can be completed to conform to the size of an actual image sensor chip.
As another example of CSP, an image sensor package proposed by the present applicant shown in FIGS. 3 and 4.
The image sensor package 40 of FIG. 3 comprises a glass substrate 41, metal wires 44 formed on the glass substrate 41, an insulating film 45 for protecting the metal wires 44, an image sensor chip 42 electrically connected to the glass substrate 41 by flipchip solder joints 43, and connection terminals 47 such as solder balls formed outside the image sensor chip 42 and connected to a printed circuit board. Meanwhile, a dust-seal layer 46 is formed between the glass substrate 41 and the image sensor chip 42 to prevent foreign substances from being introduced into a space defined between the glass substrate 41 and the image sensor chip 42.
An image sensor package 50 for a camera module shown in FIG. 4 comprises a glass substrate 51, metal wires 54 formed on the glass substrate 51, an insulating film 55 for protecting the metal wires 54, an image sensor chip 52 electrically connected to the glass substrate 51 by flipchip solder joints 53, and passive elements 58 and connection terminals 57 mounted on the metal wires 54 outside the image sensor chip 52. Although the image sensor package 50 shown in FIG. 4 has a structure nearly similar to that of the image sensor package 40 shown in FIG. 3, it has a structure in which the passive elements 58, such as decoupling capacitors, required to construct a camera module can be mounted together on the glass substrate and the connection terminals 57 for connection with a printed circuit board are provided on one surface of the glass substrate. Therefore, in case of such an image sensor package, it is basically possible to eliminate a printed circuit board in fabricating a camera module.
An image sensor package is sold as a single component for use in fabricating a camera module, or at least assembled into a camera module on a different fabrication line. That is, an image sensor package is transferred as a separate component to another line or factory and then mounted on a PCB, a flexible printed circuit (FPC) is then attached to the PCB, and a holder and a lens housing are then installed on the PCB, thereby completing a camera module. At this time, the image sensor package 20, 30, 40 or 50 is electrically connected to the PCB via the connection terminals 27, 37, 47 or 57 formed on the bottom thereof. The holder and the lens housing are installed on the PCB to surround the image sensor package 20, 30, 40 or 50, and an JR filter and a lens are installed in the holder and the lens housing such that they are located on the image sensor package.
Generally, the most critical and frequent defect in a camera module is a defect of an image sensor, which is caused by a defect of an image sensor chip itself or introduction of dust into an image sensing area during a process of packing the image sensor chip. That is, if dust particles are introduced into the image sensor package and then stick on the image sensing area, repeatable defects occur in photographed images. Even though dust particles do not stick on the image sensing area, dust molecules moving in the image sensing area are not acceptable because they may cause defects in a non-repeatable manner. Therefore, introduction of dust particles into a package or contamination of the package should be minimized during the process of packaging an image sensor. This is the reason why a line for manufacturing an image sensor package is managed at a higher degree of cleanness than lines for manufacturing other general packages.
It is known that introduction of moisture into the image sensing area degrades the color filter or micro-lens on the image sensor chip. Of course, since it takes much time for such degradation caused by moisture to appear as deterioration in image quality, it does not generally cause troubles. However, in case of products such as digital cameras for experts, which require no change in image quality for ten years or more, there is a need for a package structure capable of minimizing even the introduction of moisture.
In order to determine a defect of an image sensor itself and a defect caused by introduction of dust particles until a camera module is completed, a testing process is essential. Generally, after sensor manufacturers fabricate image sensor wafers, they perform an open and short test, and a probe test for examining whether each pixel operates properly, and then deliver map files, which show whether the sensor wafers and individual sensor chips are defective, to image sensor package manufacturers or camera module manufacturers.
The image sensor package manufacturers perform packaging of image sensors on the basis of the map files delivered from the sensor manufacturers. At this time, since a defect may be caused in an image sensor package by an error in the packaging process or introduction of dust particles, the image sensor package manufacturers perform tests for respective image sensor packages and then deliver them to the camera module manufacturers.
To complete camera modules, such test processes of determining whether image sensors are defective should also be performed. In this case, a conventional test apparatus is constructed to individually test whether image sensors are defective in the finished camera modules that have been subjected to division into separate PCB units and the process of bonding a connection means such as FPC. Therefore, after one camera module has been tested, it is pulled out manually or automatically. Subsequently, another camera module is manually or automatically seated again at a test position and then tested. Theses test procedures should be repeatedly performed. Since this method inevitably has low throughput per unit time, this becomes a factor that greatly reduces overall productivity of camera modules.
As described above, since the most critical and frequent defect in a camera module is a defect caused by introduction of dust into a pixel area of an image sensor, it is not desirable to perform a test process of determining whether an image sensor is defective after an image sensor package has been already assembled into the camera module. That is, it is desirable to determine whether an image sensor package is defective, before it is assembled into a camera module. However, since conventional apparatuses for testing existing CLCC or CSP made by Shellcase, Inc. have the function of testing whether there is a simple electrical defect by connecting the connection terminals 27 or 37 of the image sensor package 20 or 30 to external terminals and applying an electric current to the terminals, it is impossible to perform an image test that is considered to be most significant by camera module manufacturers.